The study was performed to explore the evidence for a noninfectious inflammatory cause for epidural fever. A secondary analysis of a prospective randomized trial was performed. At epidural placement, afebrile term nulliparous women were randomized to receive acetaminophen 650 mg or placebo every 4 hours. Maternal serum was collected every 4 hours until delivery. Cord blood samples were collected at delivery. Interleukin-6 (IL-6) and interleukin-8 levels were measured using enzyme-linked immunosorbent assay techniques. Student t tests, chi(2), repeated measure analysis of variance, Pearson correlation coefficients, and linear regression modeling were used as appropriate. Twenty-one subjects received placebo, and 21 received acetaminophen. The rate of fever more than 100.4 degrees F was identical in the placebo and acetaminophen groups (23.8%). Maternal serum IL-6 levels before delivery were significantly higher in mothers who had a fever (596.0 vs 169.1 pg/mL, P <.001), as was the cord blood IL-6 levels of their infants (370.5 vs 99.0 pg/mL, P <.01). Linear regression modeling demonstrated that initial maternal serum IL-6, fever, and duration of epidural but not length of rupture of membranes or number of vaginal examinations were significantly associated with final maternal serum IL-6 levels. All neonatal blood cultures were negative. Epidural fever is associated with maternal and fetal inflammation in the absence of neonatal infection. Higher levels of cytokines in maternal serum suggest that the maternal compartment is the primary inflammatory source.